The ABC Standard.

Exactness, not applause

An AI model that misses one letter of English A–Z has not produced the complete alphabet. The same absolute standard applies to every evaluated operating alphabet.

Full mastery requires recall = 1 and precision = 1. A rounded display score of 26.0 alone is insufficient evidence. The board mean is not a language-specific PASS.

Inventory sizes in v1.1

Evidence and limits

Evaluation rows may be excluded or unmeasurable. They are not silently set to zero. The supplied per-cell evidence export covers fewer models than the board; complete original evaluation traces are not provided for every model in this build.

Differences in training exposure and explicit declarations must be included in interpretations of English, Pinyin and Bantu results.